There is currently little solid information on either the prevalence or the risk factors for alcohol use among children in the U.S. Nearly all of the research has been carried out on junior high school or older samples. It is critical, however, to determine why some children begin to drink at such young ages because of the relation of problem drinking and alcoholism in later life to earlier onset of drinking. The proposed research is focused on determination of the risk factors for the childhood initiation of alcohol use, and examination of the influence of family factors in middle childhood on the development of personality and peer risk factors for later alcohol use. Early risk characteristics should be especially useful for predicting which children will move into problematic drinking as they move into middle and late adolescence. Determination of the linkages between parenting practices and the development of personality and peer risk factors foe drinking should permit the design of more effective preventive interventions. The present research focuses on two cohorts of children who will progress into preadolescence and early adolescence during this study. At the same time, they are moving from a period of low risk into a period of high risk for the initiation of alcohol use. The baseline assessments will occur before most of the children have tried alcohol outside the home. By the end of this grant period, nearly half of the older cohort will have initiated alcohol use. The research is therefore positioned to observe the transition of these children into experimentation and more regular patterns of drinking. A 7-wave, accelerated longitudinal design will be used in which children will be assessed every 6 months. Cohorts will consist of 400 White and African American families with a child aged 8 or 10 at baseline, and who will be followed up to ages 11 and 13. Families will be randomly selected from the greater Pittsburgh metropolitan area using targeted-age directory sampling and random digit dialing. African-American families will be over-sampled. An older sibling will be recruited (if available). Due to the importance of peer influence on the initiation of problem behavior, and in recognition of the problems with child perceptions of peer behavior, tow close friends will also be recruited for each participating child. Interviews, family interaction observations, and friend self-reported data will be collected during home visits. Family, peer personality, and behavioral risk factors will be constituted using these multimethod data. Measures are drawn from Problem Behavior Theory as augmented for the study of childhood socialization.